Ranger and Knox are complimentary. Knox is for perimeter security. It allows you to control the entry point for users to your cluster. You can put Knox behind a load balancer and shield the users from access to specific servers in the cluster. Ranger and Knox integrate well together, so you can use Ranger to grant permissions to users for Knox. This tutorial can walk you through setting up Knox: http://hortonworks.com/hadoop-tutorial/securing-hadoop-infrastructure-apache-knox/ And here is some good info on Knox: http://hortonworks.com/apache/knox-gateway/ ... View more

The LOCAL keyword tells Hive to write the data to the local filesystem, not HDFS. ... View more

I believe this is because you have designated your Hive table as EXTERNAL. External tables are most often used to manage data directly on HDFS that is loaded as CSV files, etc. In your case, you are creating a table and storing as ORC while populating it via a SELECT clause. Is there a reason you need it to be external? If not, you can omit the "EXTERNAL" in your "CREATE TABLE" clause and remove the "Location" entry. The only real change for you is the data will be stored under /apps/hive/warehouse on HDFS instead of the location you specified. This is what you would have: CREATE TABLE IF NOT EXISTS mi_cliente_fmes( id_interno_pe bigint, cod_nrbe_en int, mi_nom_cliente string, fec_ncto_const_pe string, fecha_prim_rl_cl string , sexo_in string, cod_est_civil_indv string, cod_est_lab_indv string, num_hijos_in int, ind_autnmo_in string, cod_ofcna_corr string, cod_cpcdad_lgl_in int ) CLUSTERED BY (cod_nrbe_en) INTO 60 BUCKETS stored as ORC ... View more